Navigate

Quick find

Navigate NeuroForge

Esc

Start typing to search public pages and evidence.

    NeuroForge open source · RKC

    Your source.
    Made understandable.

    Turn a folder, repository or documentation collection into a searchable map—with evidence you can follow.

    • Runs locally
    • No model required
    • Apache 2.0
    FoldersRepositoriesWiki exports
    RKCRepository Knowledge Compiler
    One cited atlas.People · Agents · Code

    See what is there

    An interface for
    finding your way.

    Move from an overview to a search result, a relationship or the source behind it. Keep the useful parts in view.

    RKC local workbench displaying its compiled public Go, Python and TypeScript examples, source coverage and navigation for search, graph and outputs.
    Actual RKC interface · public Go, Python and TypeScript examples

    From the big picture
    to the source.

    Start with the overview. Search for a symbol, inspect a connection or follow the evidence back to a file. Coverage and diagnostics stay close by.

    Run rkc gui to choose a local folder or search GitHub. Public repositories need no account connection. For private repositories, connect with a personal access token held only for your local server session.

    Use it your way

    One foundation.
    Three ways in.

    Use the browser, automate with scripts or connect an agent.

    For people

    Search source, follow relationships and inspect evidence in a local browser.

    For agents

    Give agents bounded, cited context through MCP or HTTP. Keep source identity and coverage with each result.

    For code

    Automate scans and integrate versioned records using the command line or typed Go client.

    Useful beyond the scan

    Take the knowledge
    with you.

    Use one repository or combine processed folders, documentation and local wiki exports.

    A map you can explore

    Files, symbols and relationships, with diagnostics showing where analysis is incomplete.

    Context you can trace

    Markdown or JSON evidence packets for a question, cited back to the compiled source.

    Knowledge you can move

    Portable packs preserving content hashes, provenance and source licence information.

    Start with your own source

    Make your first atlas.

    1. Install RKC, then run rkc gui.
    2. Choose a folder or GitHub repository.
    3. Select Compile. Explore your atlas in the same window.

    Portable downloads need no Go, Python or model. Folder compilation stays local; GitHub acquisition connects to GitHub.

    Portable downloads are available. See the release receipts for tested platforms. Release details

    Requires user-systemd and delegated cgroup v2 controllers for compilation. Installer: curl, unzip and sha256sum or shasum.

    1. Install · Terminal
    curl -fsSL https://github.com/neuroforge-io/RKC/releases/latest/download/install-release.sh | sh
    2. Open RKCTerminal
    export PATH="$HOME/.local/bin:$PATH"
    rkc gui

    Built to be useful. Open to build on.

    A public contribution
    from NeuroForge.

    Add an analyzer, build an integration, report a rough edge or use RKC in a project of your own.

    Original RKC code is available under Apache 2.0. Source material and third-party components keep their own licences.

    RKC is an independent open-source project from NeuroForge. ERAIS is a separate proprietary system.